Characterization of the clinical effects after the first dose of bacterially synthesized recombinant human granulocyte-macrophage colony-stimulating factor

Bacterially synthesized recombinant human granulocyte-macrophage colony-stimulating factor (rhGM-CSF) is an agent with therapeutic potential for neutropenic states, but even at doses below the maximal tolerated dose adverse effects occur during short courses of administration. We have recognized a syndrome of hypoxia and hypotension that follows the first but not subsequent doses of rhGM-CSF. Thirteen of 42 patients receiving rhGM-CSF in phase I studies an 4 of 6 patients in a phase II study developed a reaction that occurred after the first dose of 24 of 78 cycles of rhGM-CSF therapy. The reaction was characterized by flushing (16 of 24), tachycardia (16 of 24), hypotension (14 of 24), musculoskeletal pain (13 of 24), dyspnea (12 of 24), nausea and vomiting (11 of 24), rigors (5 of 24), involuntary leg spasms (3 of 24), and syncope (3 of 24). The reaction did not occur after any of more than 600 second and subsequent consecutive rhGM-CSF doses. Oxygen saturation decreased during first-dose reactions by 8% ± 4% as compared with 3% ± 1% on first days without reactions (P < .001). Pulmonary dysfunction was characterized by hypoxemia (59 ± 9 mm Hg, mean ± SD) that was fully correctable with supplementary oxygen, decreased single-breath carbon monoxide diffusion capacity, and increased alveolar-arterial oxygen gradients (25 ± 6 to 60 ± 4 mm Hg, mean ± SD), but no significant abnormalities on chest roentgenogram or lung perfusion scan. Factors predisposing to reactions were rhGM-CSF dose ≥ 3 μg/kg (P < .01), intravenous (IV) rather than subcutaneous (SC) administration (P < .05), occurrence of a reaction after the first dose of a previous cycle of rhGM-CSF therapy (P < .01), and for patients receiving 15 μg/kg/d by SC bolus, the presence of lung cancer (P < .05). Administration of 15 μg/kg/d rhGM-CSF by 24-hour SC infusion rather than SC bolus resulted in a delayed onset of reaction from 30 ± 8 minutes to 240 ± 190 minutes (mean ± SD, P < .001), and a slower rate of initial transient decrease in neutrophil levels and a more prolonged duration of transient leukopenia. The time of onset of reactions correlated with he rate of rise of rhGM-CSF levels. After rhGM-CSF there was no change in serum levels of total hemolytic complement, the third and fourth component of complement, histamine, and tumor necrosis factor-α (TNF-α). We conclude that transient hypoxia and hypotension are significant complications of the first dose of rhGM-CSF that are not solely explained by the transient leukopenia or pulmonary sequestration of leukocytes. The cardiovascular changes and rise in alveolar-arterial oxygen gradient suggest that regional intrapulmonary ventilation/perfusion mismatching owing to release of a vasoactive mediator is a significant contributing factor.